Climate Change Migration Network
Fondation Jean-Jacques Laffont - Toulouse Sciences Economiques
Kyllian James
Creation Date: 2023-10-16 | Last Modified Date: 2023-11-21
1 Load Librairies
library(knitr)
library(readxl)
library(dplyr)
library(tidyr)
library(DT)
library(sf)
library(ggplot2)
library(ggspatial)
library(gridExtra)
library(rnaturalearth)
library(rnaturalearthhires) #GitHub: ropensci/rnaturalearthhires
library(leaflet)
library(plotly)
library(ggiraph)
library(openxlsx)
library(DescTools)
LibrariesVersions <- data.frame(
LIBRARY = c("KNITR", "READXL", "DPLYR", "TIDYR", "DT",
"SF", "GGPLOT2", "GGSPATIAL", "GRIDEXTRA", "RNATURALEARTH",
"RNATURALEARTHHIRES", "LEAFLET", "PLOTLY", "GGIRAPH", "OPENXLSX",
"DESCTOOLS"),
VERSION = c(as.character(packageVersion("knitr")),
as.character(packageVersion("readxl")),
as.character(packageVersion("dplyr")),
as.character(packageVersion("tidyr")),
as.character(packageVersion("DT")),
as.character(packageVersion("sf")),
as.character(packageVersion("ggplot2")),
as.character(packageVersion("ggspatial")),
as.character(packageVersion("gridExtra")),
as.character(packageVersion("rnaturalearth")),
as.character(packageVersion("rnaturalearthhires")),
as.character(packageVersion("leaflet")),
as.character(packageVersion("plotly")),
as.character(packageVersion("ggiraph")),
as.character(packageVersion("openxlsx")),
as.character(packageVersion("DescTools"))))
datatable(LibrariesVersions,
options = list(scrollX = TRUE, autoWidth = TRUE),
caption = paste("Librairies Versions | R Version:",
gsub("^R version\\s*|\\s*\\(\\d{4}-\\d{2}-\\d{2} ucrt\\)", "", R.version.string)),
style = "bootstrap", rownames = FALSE) %>%
formatStyle(columns = 1, textAlign = "left") %>%
formatStyle(columns = 2, textAlign = "center")2 ISO and M49 Codes
ISO 3166-1 is an international standard from International Organization for Standardization (ISO), with 3 codes:
- ISO2
- ISO3
- ISO.NUMERIC
ISO.NUMERIC codes are identical to M49 codes from UNSD. M49 geographical entities are not all in ISO codes. ISO 3166-1 is only available for autonomous, but not necessarily independent, countries or territories. In addition, M49 codes are defined for the world, continents and geographical groupings. How to get a code?
- Being a UN member state
- Be a member of one of the UN’s specialized agencies (e.g. FAO | ILO | IMF| UNESCO | World Bank | WHO | …)
There are more ISO.NUM/M49 codes (249) than countries recognized by the UN (198, including 193 member states and 5 observer or associate states).
3 Data
Stock data are different from flow data. The first one corresponds to the number of people in a country, while the second one is the number of people crossing a country’s borders.
International migrant stock data (2020) are taken from the United Nations (ONU | Population Division). The estimates of the number (or “stock”) of international migrants (1990-2020) are available for 232 countries/areas of the world.
The international migrant population corresponds to the total number of migrants present in a given country at a given moment in time (foreign-born). Tourists are not counted among migrants.
Data are organized by residence-birth pairs according to certain classifications or countries/areas:
- 6 Continental
Regions (M49 | ISO3) with Sub-Regions and Intermediate Regions:
- Africa
- Asia
- Europe
- Latin America and the Caribbean
- Northern America
- Oceania
- 8 Geographic Regions based on the Sustainable
Development Goals (SDG) Classification (UN):
- Sub-Saharan Africa
- Northern Africa and Western Asia
- Central and Southern Asia
- Eastern and South-Eastern Asia
- Latin America and the Caribbean
- Oceania (excluding Australia and New Zealand)
- Australia and New Zealand
- Europe and Northern America
- 2 Development Levels:
- More Developed Regions:
- Europe
- Northern America
- Australia and New Zealand and Japan
- Less Developed Regions:
- Africa
- Asia (excluding Japan)
- Latin America and the Caribbean
- Oceania (excluding Australia and New Zealand)
- More Developed Regions:
- 3 Development Levels (137/232):
- Least Developed Countries (LDCs | 47)
- Land-Locked Developing Countries (LLDCs | 32)
- Small Island Developing States (SIDS | 58)
- 3/4 Income Levels based on Gross National Income (GNI | World
Bank 2020):
- High-Income
- Middle-Income
- Upper-Middle-Income
- Lower-Middle-Income
- Low-Income
These income groups are not available for all countries/areas.
ResidenceCode:
- M49: Regions
- ISO.NUMERIC: Countries
3.1 Migration Stock Data
International Migrant Stocks (IMS) in the World
(Residence) born in Europe and Northern America
(Birth) are shown in this table. We have estimates from
1990 to 2020 (Year):
IMSTis IMS (Both Sexes Combined)IMSMis IMS (Males)IMSFis IMS (Females)
imst <- read_xlsx("../Données/IMS/ONU/Population Division/undesa_pd_2020_ims_stock_by_sex_destination_and_origin.xlsx",
sheet = "Table 1", skip = 8) %>%
select(c(2, 4, 6:14)) %>%
rename(Residence = 1, ResidenceCode = 2, Birth = 3, BirthCode = 4,
`1990` = 5, `1995` = 6, `2000` = 7, `2005` = 8, `2010` = 9, `2015` = 10, `2020` = 11)
imst <- pivot_longer(imst, cols = 5:11) %>% rename(Year = 5, IMST = 6)
imsm <- read_xlsx("../Données/IMS/ONU/Population Division/undesa_pd_2020_ims_stock_by_sex_destination_and_origin.xlsx",
sheet = "Table 1", skip = 8) %>%
select(c(2, 4, 6:7, 15:21)) %>%
rename(Residence = 1, ResidenceCode = 2, Birth = 3, BirthCode = 4,
`1990` = 5, `1995` = 6, `2000` = 7, `2005` = 8, `2010` = 9, `2015` = 10, `2020` = 11)
imsm <- pivot_longer(imsm, cols = 5:11) %>% rename(Year = 5, IMSM = 6)
imsf <- read_xlsx("../Données/IMS/ONU/Population Division/undesa_pd_2020_ims_stock_by_sex_destination_and_origin.xlsx",
sheet = "Table 1", skip = 8) %>%
select(c(2, 4, 6:7, 22:28)) %>%
rename(Residence = 1, ResidenceCode = 2, Birth = 3, BirthCode = 4,
`1990` = 5, `1995` = 6, `2000` = 7, `2005` = 8, `2010` = 9, `2015` = 10, `2020` = 11)
imsf <- pivot_longer(imsf, cols = 5:11) %>% rename(Year = 5, IMSF = 6)
ims <- cbind(imst, imsm[length(imsm)], imsf[length(imsf)])
kable(ims[57:63, c(1, 3, 5:8)], row.names = FALSE,
align = c("l", "l", "c", "c", "c", "c"),
caption = "World's Migration Stocks born in Europe and Northern America (1990-2020)")| Residence | Birth | Year | IMST | IMSM | IMSF |
|---|---|---|---|---|---|
| WORLD | Europe and Northern America | 1990 | 50532413 | 23704534 | 26827879 |
| WORLD | Europe and Northern America | 1995 | 51582164 | 24148758 | 27433406 |
| WORLD | Europe and Northern America | 2000 | 52877703 | 24676104 | 28201599 |
| WORLD | Europe and Northern America | 2005 | 54849090 | 25682798 | 29166292 |
| WORLD | Europe and Northern America | 2010 | 58681104 | 27337252 | 31343852 |
| WORLD | Europe and Northern America | 2015 | 60592402 | 28229706 | 32362696 |
| WORLD | Europe and Northern America | 2020 | 67601621 | 31770385 | 35831236 |
This was an example between World-(Europe and NA). Data are also available for more regions/countries (232).
3.2 Geo Data
We will need World Administrative Boundaries (WAB) to build maps. WAB are available on OpenDataSoft (GeoJSON file is used). The Coordinate Reference System ESRI:54030 corresponds to the World Robinson Projection.
WAB <- read_sf("../Données/Geo/OpenDataSoft/world_administrative_boundaries.geojson")
ESRI_54030 <- "+proj=robin +lon_0=0 +x_0=0 +y_0=0 +ellps=WGS84 +datum=WGS84 +units=m +no_defs"
WAB <- st_transform(WAB, ESRI_54030)
WAB <- WAB %>% select(-1) %>%
rename(ISO3 = iso3, Status = status, ColorCode = color_code, Name = name, Continent = continent, Region = region, ISO2 = iso_3166_1_alpha_2_codes, FName = french_short)
GGPLOT_WAB <- ggplot() +
geom_sf(data = WAB, aes(fill = ISO3)) +
labs(title = "World Administrative Boundaries (Source: OpenDataSoft)",
x = "Longitude", y = "Latitude") +
theme(legend.position = 'none') +
ggspatial::annotation_north_arrow(location = "tr", which_north = TRUE,
height = unit(1, "cm"), width = unit(1, "cm"),
style = north_arrow_fancy_orienteering(
line_col = NA,
fill = c("black", "black"),
text_size = 5))
GGPLOT_WAB
There are 256 Name | 237 ISO3 | 207
ColorCode.
There is a variable Continent:
- Africa
- Americas
- Antartica
- Asia
- Europe
- Oceania
There is also variable Region:
WAB$Status <- sub("Sovereignty unsettled", "Sovereignty Unsettled", WAB$Status)
WAB$Status <- sub("Sovereignty unsettled JammuK-China", "Sovereignty Unsettled JammuK-China", WAB$Status)
WAB_NoGeo <- WAB %>% st_drop_geometry()
datatable(WAB_NoGeo %>% group_by(Region) %>% summarise(Count = n()),
options = list(scrollX = TRUE, autoWidth = TRUE,
columnDefs = list(list(targets = "_all", render = JS(
"function(data, type, row, meta) {",
" if (type === 'display' && data === null) {",
" return 'NA';",
" }",
" return data;",
"}")))),
style = "bootstrap", rownames = FALSE) %>%
formatStyle(columns = 1, textAlign = "left") %>%
formatStyle(columns = 2, textAlign = "center")Some changes are made in Region for South Sudan to match
with UNSD information used later.
WAB[which(WAB$ISO3 == "SSD"), "Region"] <- "Eastern Africa"We would like to be able to represent every Residence
that is in migration stock data.
ResidencesIMS <- unique(ims[c("Residence", "ResidenceCode")])
ResidencesIMS <- ResidencesIMS[ResidencesIMS$ResidenceCode < 900, ]
ResidencesIMS$Residence <- gsub("\\*", "", ResidencesIMS$Residence)
datatable(ResidencesIMS, options = list(scrollX = TRUE, autoWidth = TRUE),
caption = "Residences in Migration Stock Data",
style = "bootstrap", rownames = FALSE) %>%
formatStyle(columns = 1, textAlign = "left") %>%
formatStyle(columns = 2, textAlign = "center")There are 235 Residence (\(\ne\) 256 Name in WAB | \(\simeq\) 237 ISO3 in WAB).
datatable(WAB_NoGeo %>% group_by(Status) %>% summarise(Count = n()),
options = list(scrollX = TRUE, autoWidth = TRUE),
style = "bootstrap", rownames = FALSE) %>%
formatStyle(columns = 1, textAlign = "left") %>%
formatStyle(columns = 2, textAlign = "center")There are 193/256 member states and 63/256 areas/territories which
are not member states. This could explain why there are more
ISO3 than member states.
First, we will combine ResidencesIMS and WAB according to
Residence and Name to see if there are any
matching problems. Then, we will explore WAB.
RIMS_WAB <- left_join(ResidencesIMS, WAB, by = c("Residence" = "Name"))
RIMS_WAB_NoGeo <- RIMS_WAB %>% select(-ncol(RIMS_WAB))
datatable(RIMS_WAB_NoGeo[is.na(RIMS_WAB_NoGeo$ISO3), ][ , 1:3],
options = list(scrollX = TRUE, autoWidth = TRUE,
columnDefs = list(list(targets = "_all", render = JS(
"function(data, type, row, meta) {",
" if (type === 'display' && data === null) {",
" return 'NA';",
" }",
" return data;",
"}")))),
style = "bootstrap", rownames = FALSE,
caption = "Non-Matching Rows from WAB (R) in ResidencesIMS (L)") %>%
formatStyle(columns = 1, textAlign = "left") %>%
formatStyle(columns = c(2:3), textAlign = "center")WAB_RIMS <- left_join(WAB, ResidencesIMS, by = c("Name" = "Residence"))
WAB_RIMS_NoGeo <- WAB_RIMS %>% st_drop_geometry()
datatable(WAB_RIMS_NoGeo[is.na(WAB_RIMS_NoGeo$ResidenceCode), ][ , c(4, 9)],
options = list(scrollX = TRUE, autoWidth = TRUE,
columnDefs = list(list(targets = "_all", render = JS(
"function(data, type, row, meta) {",
" if (type === 'display' && data === null) {",
" return 'NA';",
" }",
" return data;",
"}")))),
style = "bootstrap", rownames = FALSE,
caption = "Non-Matching Rows from ResidencesIMS (R) in WAB (L)") %>%
formatStyle(columns = 1, textAlign = "left") %>%
formatStyle(columns = 2, textAlign = "center")Residence \(\neq\)
Name causes lines-matching problems, we need to change
Residence in ResidencesIMS or Name in WAB.
Jersey and Guernsey in WAB correspond to the Channel Islands in
ResidencesIMS, while West Bank and Gaza Strip correspond to State of
Palestine.
Netherlands Antilles in WAB includes Bonaire and Curaçao. There are 3
Residence in ResidencesIMS:
- Bonaire, Sint Eustatius and Saba
- Sint Marteen (Dutch Part)
- Curaçao
We find missing geometries in rnaturalearth: Saba, Sint
Eustatius and Sint Marteen.
There are 5 Residence in ResidencesIMS without
geometries in WAB:
- Saint Helena
- Saint Barthélemy
- Saint Martin (French Part)
- Saint Pierre and Miquelon
- Wallis and Futuna Islands
We find these missing geometries also in
rnaturalearth.
Lines-matching problems from WAB (R) in ResidencesIMS (L) are now solved. There are still lines-matching problems from ResidencesIMS (R) in WAB (L).
WAB[WAB$Name == "Reunion", "Name"] <- "Réunion"
WAB[WAB$Name == "Libyan Arab Jamahiriya", "Name"] <- "Libya"
WAB[WAB$Name == "Swaziland", "Name"] <- "Eswatini"
WAB[WAB$Name == "Cape Verde", "Name"] <- "Cabo Verde"
WAB[WAB$Name == "Hong Kong", "Name"] <- "China, Hong Kong SAR"
WAB[WAB$Name == "Macao", "Name"] <- "China, Macao SAR"
WAB[WAB$Name == "Taiwan", "Name"] <- "China, Taiwan Province of China"
WAB[WAB$Name == "Democratic People's Republic of Korea", "Name"] <- "Dem. People's Republic of Korea"
WAB[WAB$Name == "Vietnam", "Name"] <- "Viet Nam"
WAB[WAB$Name == "Czech Republic", "Name"] <- "Czechia"
WAB[WAB$Name == "Moldova, Republic of", "Name"] <- "Republic of Moldova"
WAB[WAB$Name == "U.K. of Great Britain and Northern Ireland", "Name"] <- "United Kingdom"
WAB[WAB$Name == "Bosnia & Herzegovina", "Name"] <- "Bosnia and Herzegovina"
WAB[WAB$Name == "The former Yugoslav Republic of Macedonia", "Name"] <- "North Macedonia"
WAB[WAB$Name == "Antigua & Barbuda", "Name"] <- "Antigua and Barbuda"
WAB[WAB$Name == "Bolivia", "Name"] <- "Bolivia (Plurinational State of)"
WAB[WAB$Name == "Venezuela", "Name"] <- "Venezuela (Bolivarian Republic of)"
WAB[WAB$Name == "Micronesia (Federated States of)", "Name"] <- "Micronesia (Fed. States of)"
WestBank <- WAB %>% filter(ISO3 == "PSE" & Name == "West Bank")
GazaStrip <- WAB %>% filter(Name == "Gaza Strip")
GeoPSE <- st_union(WestBank, GazaStrip) %>% select(c(1:8, 17))
WAB[which(WAB$ISO3 == "ANT"), "ISO2"] <- "AN"
WAB[which(WAB$ISO3 == "IMY"), "ISO2"] <- "IM"
NA_in_ISO3 <- WAB[is.na(WAB$ISO3), ]
WAB <- WAB %>% filter(ISO3 != "PSE") %>% rbind(GeoPSE)
WAB[which(WAB$ISO3 == "PSE"), "ColorCode"] <- "PSE"
WAB[which(WAB$ISO3 == "PSE"), "Name"] <- "State of Palestine"
Jersey <- NA_in_ISO3 %>% filter(Name == "Jersey")
Guernsey <- NA_in_ISO3 %>% filter(Name == "Guernsey")
GeoCHA <- st_union(Jersey, Guernsey) %>% select(c(1:8, 17))
NA_in_ISO3 <- NA_in_ISO3 %>% filter(!(Name %in% c("Jersey", "Guernsey")))
WAB <- WAB %>% rbind(GeoCHA)
WAB[which(WAB$Name == "Jersey"), "ISO3"] <- "CHA"
WAB[which(WAB$ISO3 == "CHA"), "ColorCode"] <- "CHA"
WAB[which(WAB$ISO3 == "CHA"), "Name"] <- "Channel Islands"
WAB[which(WAB$ISO3 == "CHA"), "FName"] <- "Îles Anglo-Normandes"
NL_Antilles <- WAB %>% filter(Name == "Netherlands Antilles")
NL_Antilles <- st_cast(NL_Antilles, "POLYGON")
NL_Antilles <- NL_Antilles %>%
mutate(Name = c("Bonaire", "Curaçao"),
ISO3 = c("BES", "CUW"),
ColorCode = c("BES", "CUW"),
ISO2 = c("BQ", "CW"),
FName = c("Bonaire", "Curaçao"))
WAB <- WAB %>% filter(ISO3 != "ANT") %>% rbind(NL_Antilles)
WAB_RNE <- ne_states(returnclass = "sf")
WAB_RNE <- st_transform(WAB_RNE, ESRI_54030)
SintMarteen <- WAB_RNE %>% filter(adm0_a3 == "SXM")
SintMarteenGeo <- st_geometry(SintMarteen)
SintMarteen <- st_sf(geometry = SintMarteenGeo, crs = ESRI_54030)
WAB_RNEw <- SintMarteen %>%
mutate(ISO3 = "SXM",
Status = "NL Territory",
ColorCode = "SXM",
Name = "Sint Maarten (Dutch part)",
Continent = "Americas",
Region = "Caribbean",
ISO2 = "SX",
FName = "Saint-Martin (partie néerlandaise)") %>%
select(-1, everything())
SintEustatiusSaba <- WAB_RNE %>% filter(name %in% c("St. Eustatius", "Saba"))
SintEustatiusSabaGeo <- st_geometry(SintEustatiusSaba)
SintEustatiusSaba <- st_sf(geometry = SintEustatiusSabaGeo, crs = ESRI_54030)
SintEustatiusSaba <- SintEustatiusSaba %>%
mutate(ISO3 = c("BES", "BES"),
Status = c("NL Territory", "NL Territory"),
ColorCode = c("BES", "BES"),
Name = c("Sint Eustatius", "Saba"),
Continent = c("Americas", "Americas"),
Region = c("Caribbean", "Caribbean"),
ISO2 = c("BQ", "BQ"),
FName = c("Saint-Eustache", "Saba")) %>%
select(-1, everything())
Bonaire <- WAB %>% filter(ISO3 == "BES")
GeoBES <- st_union(Bonaire, st_union(SintEustatiusSaba))
WAB <- WAB %>% filter(ISO3 != "BES")
GeoBES <- GeoBES %>%
mutate(Name = "Bonaire, Sint Eustatius and Saba",
FName = "Bonaire, Saint-Eustache and Saba")
WAB_RNEw <- rbind(WAB_RNEw, GeoBES)
HelenaBarthMartin <- WAB_RNE %>%
filter(name %in% c("Saint Helena", "Saint Barthélemy", "St. Martin"))
HelenaBarthMartinGeo <- st_geometry(HelenaBarthMartin)
HelenaBarthMartin <- st_sf(geometry = HelenaBarthMartinGeo, crs = ESRI_54030)
HelenaBarthMartin <- HelenaBarthMartin %>%
mutate(ISO3 = c("MAF", "BLM", "SHN"),
Status = c("FR Territory", "FR Territory", "UK Territory"),
ColorCode = c("MAF", "BLM", "SHN"),
Name = c("Saint Martin (French part)", "Saint Barthélemy", "Saint Helena"),
Continent = c("Americas", "Americas", "Africa"),
Region = c("Caribbean", "Caribbean", "Western Africa"),
ISO2 = c("MF", "BL", "SH"),
FName = c("Saint-Martin (partie française)", "Saint-Barthélemy", "Sainte-Hélène")) %>%
select(-1, everything())
WAB_RNEw <- rbind(WAB_RNEw, HelenaBarthMartin)
SaintPierreMiquelon <- WAB_RNE %>% filter(adm0_a3 == "SPM")
SaintPierreMiquelonGeo <- st_geometry(SaintPierreMiquelon)
SaintPierreMiquelonGeo <- st_union(SaintPierreMiquelonGeo)
SaintPierreMiquelon <- st_sf(geometry = SaintPierreMiquelonGeo, crs = ESRI_54030) %>%
mutate(ISO3 = "SPM",
Status = "FR Territory",
ColorCode = "SPM",
Name = "Saint Pierre and Miquelon",
Continent = "Americas",
Region = "Northern America",
ISO2 = "PM",
FName = "Saint-Pierre-et-Miquelon") %>%
select(-1, everything())
WAB_RNEw <- rbind(WAB_RNEw, SaintPierreMiquelon)
WallisFutuna <- WAB_RNE %>% filter(adm0_a3 == "WLF")
WallisFutunaGeo <- st_geometry(WallisFutuna)
WallisFutunaGeo <- st_union(WallisFutunaGeo)
WallisFutuna <- st_sf(geometry = WallisFutunaGeo, crs = ESRI_54030) %>%
mutate(ISO3 = "WLF",
Status = "FR Territory",
ColorCode = "WLF",
Name = "Wallis and Futuna Islands",
Continent = "Oceania",
Region = "Polynesia",
ISO2 = "WF",
FName = "Wallis-et-Futuna (Îles)") %>%
select(-1, everything())
WAB_RNEw <- rbind(WAB_RNEw, WallisFutuna)
WAB <- rbind(WAB, WAB_RNEw)
WAB <- WAB %>% rbind(NA_in_ISO3)
RIMS_WAB <- left_join(ResidencesIMS, WAB, by = c("Residence" = "Name"))
RIMS_WAB_NoGeo <- RIMS_WAB %>% select(-ncol(RIMS_WAB))
WAB_RIMS <- left_join(WAB, ResidencesIMS, by = c("Name" = "Residence"))
WAB_RIMS_NoGeo <- WAB_RIMS %>% st_drop_geometry()
datatable(WAB_RIMS_NoGeo[is.na(WAB_RIMS_NoGeo$ResidenceCode), ][ , c(4, 9)],
options = list(scrollX = TRUE, autoWidth = TRUE,
columnDefs = list(list(targets = "_all", render = JS(
"function(data, type, row, meta) {",
" if (type === 'display' && data === null) {",
" return 'NA';",
" }",
" return data;",
"}")))),
style = "bootstrap", rownames = FALSE,
caption = "Non-Matching Rows from ResidencesIMS (R) in WAB (L)") %>%
formatStyle(columns = 1, textAlign = "left") %>%
formatStyle(columns = 2, textAlign = "center")These Name are mostly islands. Let’s explore WAB to
understand better why they are not in Residence.
WAB_NoGeo <- WAB %>% st_drop_geometry()
kable(t(colSums(is.na(WAB_NoGeo[1:6]))), row.names = FALSE,
align = "c", caption = "NA Values in Geo Data")| ISO3 | Status | ColorCode | Name | Continent | Region |
|---|---|---|---|---|---|
| 16 | 1 | 1 | 0 | 0 | 4 |
There is one missing value in Status and
ColorCode:
kable(WAB_NoGeo[1:4][is.na(WAB_NoGeo[1:4]$Status) & is.na(WAB_NoGeo[1:4]$ColorCode), ],
align = c("c", "c", "c", "l"))| ISO3 | Status | ColorCode | Name |
|---|---|---|---|
| NA | NA | NA | Abyei |
Abyei is an area, currently controlled by Sudan but claimed by South
Sudan, on the border between South Sudan and Sudan that has been
accorded “special administrative status” by the 2004 Protocol on the
Resolution of the Abyei Conflict (Abyei Protocol) in the Comprehensive
Peace Agreement (CPA) that ended the Second Sudanese Civil War. We
decided to keep missing values for this area and to add Abyei to
ResidencesIMS with a missing value in ResidenceCode.
ResidencesIMS <- rbind(ResidencesIMS,
data.frame(Residence = "Abyei", ResidenceCode = NA))
ResidencesIMS <- data.frame(ResidencesIMS, row.names = NULL)What about missing values in ISO3 ?
datatable(WAB_NoGeo[1:4][is.na(WAB_NoGeo[1:4]$ISO3), ],
options = list(scrollX = TRUE, autoWidth = TRUE,
columnDefs = list(list(targets = "_all", render = JS(
"function(data, type, row, meta) {",
" if (type === 'display' && data === null) {",
" return 'NA';",
" }",
" return data;",
"}")))),
style = "bootstrap", rownames = FALSE) %>%
formatStyle(columns = c(2, 4), textAlign = "left") %>%
formatStyle(columns = c(1, 3), textAlign = "center")There are 4 areas with \(XXX\) |
\(XXY\) | \(XXZ\) as ColorCode:
- Paracel Islands are de facto controlled by China, but claimed by Taiwan and Vietnam.
- Spratly Islands are occupied by military forces from Brunei, China, Malaysia, Philippines, Taiwan and Vietnam.
- Aksai Chin is a region controlled by China but claimed by India.
- Jammu-Kashmir is a region administered by India.
Sovereignty is unsettled for these 4 areas. We decided to keep
missing values in ISO3, to replace ColorCode
by missing values and to add them to ResidencesIMS with missing values
in ResidenceCode.
WAB[which(WAB$Name == "Paracel Islands"), "ColorCode"] <- NA
WAB[which(WAB$Name == "Spratly Islands"), "ColorCode"] <- NA
WAB[which(WAB$Name == "Aksai Chin"), "ColorCode"] <- NA
WAB[which(WAB$Name == "Jammu-Kashmir"), "ColorCode"] <- NA
WAB_NoGeo <- WAB %>% st_drop_geometry()
ResidencesIMS <- rbind(ResidencesIMS,
data.frame(Residence = c("Paracel Islands",
"Spratly Islands",
"Aksai Chin",
"Jammu-Kashmir"),
ResidenceCode = c(NA, NA, NA, NA)))What about the other missing values in ISO3 and
Region ?
NA_in_ISO3_Region <- WAB_NoGeo[c(1:4, 6)][is.na(WAB_NoGeo[c(1:4, 6)]$ISO3) |
is.na(WAB_NoGeo[c(1:4, 6)]$Region), ]
NA_in_ISO3_Region <- subset(NA_in_ISO3_Region, !(is.na(ISO3) & is.na(ColorCode)))
datatable(NA_in_ISO3_Region,
options = list(scrollX = TRUE, autoWidth = TRUE,
columnDefs = list(list(targets = "_all", render = JS(
"function(data, type, row, meta) {",
" if (type === 'display' && data === null) {",
" return 'NA';",
" }",
" return data;",
"}")))),
style = "bootstrap", rownames = FALSE) %>%
formatStyle(columns = c(2, 4), textAlign = "left") %>%
formatStyle(columns = c(1, 3), textAlign = "center")There are 3 areas with missing value in Region:
- French Southern and Antarctic Territories
- Bouvet Island
- Heard Island and McDonald Islands
We can combine their geometries with those of their respective
countries (sovereign state or administrator state) but as these zones
have a numeric iso
code, we decided to keep the missing values in Region
(located in Antarctica continent) and add them to ResidencesIMS with
their respective ResidenceCode values.
There are 11 areas with missing value in ISO3:
- Kuril Islands are located between Japan and Russia.
- Ma’tan al-Sarra is a military base between Egypt and Sudan.
- Glorioso Islands are located between Mayotte and Madagascar.
- Ilemi Triangle is a disputed area claimed by South Sudan and Kenya.
- Arunachal Pradesh is an Indian state.
- Guantanamo is an U.S. military detention center.
- Jarvis and Midway Islands are part of the United States Minor Outlying Islands.
- South Georgia & the South Sandwich Islands have an
ISO3. - Hala’ib Triangle is a disputed area claimed by Egypt and Sudan
- Madeira Islands are portuguese islands.
South Georgia & the South Sandwich Islands (ISO3 = SGS | ISO2 =
GS) is located in Antarctica continent. We keep missing value in
Region and add it to ResidencesIMS with
ResidenceCode = 239.
We decided to keep missing values in ISO3 and
ISO2 and to add some areas to ResidencesIMS with missing
values in ResidenceCode:
- Ma’tan al-Sarra
- Ilemi Triangle
- Guantanamo
- Hala’ib Triangle
We decided to combine some geometries with their respective member states:
- Kuril Islands with Russian Federation
- Glorioso Islands with France
- Arunachal Pradesh with India
- Madeira Islands with Portugal
For statistical purposes, certain areas are often grouped together. We will combine their geometries and add them to ResidencesIMS:
- Jarvis and Midway Islands with
ISO3= UMI |ISO2= UM |ResidenceCode= 581
Then, we combine again ResidencesIMS and WAB according to
Residence and Name to see if there are still
any matching problems.
ResidencesIMS <- rbind(ResidencesIMS,
data.frame(Residence = c("French Southern and Antarctic Territories",
"Bouvet Island",
"Heard Island and McDonald Islands",
"South Georgia & the South Sandwich Islands",
"Ma'tan al-Sarra",
"Ilemi Triangle",
"Guantanamo",
"Hala'ib Triangle"),
ResidenceCode = c(260, 74, 334, 239,
NA, NA, NA, NA)))
WAB[which(WAB$Name == "South Georgia & the South Sandwich Islands"), "ISO3"] <- "SGS"
WAB[which(WAB$Name == "South Georgia & the South Sandwich Islands"), "ISO2"] <- "GS"
WAB_NA_in_ISO3_GEOC <- WAB[is.na(WAB$ISO3) &
WAB$ColorCode %in% c("RUS", "FRA", "IND", "PRT"), ]
WAB_MbS <- WAB[!is.na(WAB$ISO3) & WAB$Status == "Member State", ]
for (i in 1:nrow(WAB_NA_in_ISO3_GEOC)) {
MATCHR <- WAB_MbS[WAB_MbS$ColorCode == WAB_NA_in_ISO3_GEOC$ColorCode[i], ]
CC <- MATCHR[3][[1]]
GEOC <- st_union(MATCHR, WAB_NA_in_ISO3_GEOC[i, ])
WAB <- WAB %>% filter(Name != WAB_NA_in_ISO3_GEOC[i, 4][[1]])
WAB[which(WAB$Name == GEOC$Name), ]$geometry <- GEOC$geometry
}
GeoUMI <- WAB[WAB$Name %in% c("Jarvis Island", "Midway Is."), ]
GeoUMI <- st_union(GeoUMI)
UMI <- st_sf(geometry = GeoUMI, crs = ESRI_54030)
UMI <- UMI %>%
mutate(ISO3 = "UMI",
Status = "US Territory",
ColorCode = "UMI",
Name = "United States Minor Outlying Islands",
Continent = "Americas",
Region = "Northern America",
ISO2 = "UM",
FName = "Îles mineures éloignées des États-Unis") %>%
select(-1, everything())
WAB <- WAB %>% filter(!Name %in% c("Jarvis Island", "Midway Is."))
WAB <- rbind(WAB, UMI)
ResidencesIMS <- rbind(ResidencesIMS,
data.frame(Residence = "United States Minor Outlying Islands",
ResidenceCode = 581))
RIMS_WAB <- left_join(ResidencesIMS, WAB, by = c("Residence" = "Name"))
RIMS_WAB_NoGeo <- RIMS_WAB %>% select(-ncol(RIMS_WAB))
WAB_RIMS <- left_join(WAB, ResidencesIMS, by = c("Name" = "Residence"))
WAB_RIMS_NoGeo <- WAB_RIMS %>% st_drop_geometry()
datatable(WAB_RIMS_NoGeo[is.na(WAB_RIMS_NoGeo$ResidenceCode) &
!is.na(WAB_RIMS_NoGeo$ISO3), ][ , c(1:4, 9)],
options = list(scrollX = TRUE, autoWidth = TRUE,
columnDefs = list(list(targets = "_all", render = JS(
"function(data, type, row, meta) {",
" if (type === 'display' && data === null) {",
" return 'NA';",
" }",
" return data;",
"}")))),
style = "bootstrap", rownames = FALSE,
caption = "Non-Matching Rows from ResidencesIMS (R) in WAB (L)") %>%
formatStyle(columns = c(2, 4), textAlign = "left") %>%
formatStyle(columns = c(1, 3, 5), textAlign = "center")There are 7 areas that have an ISO3 but are not in
ResidencesIMS. Each of them except Azores Islands have a numeric iso
code, so we decided to add them to ResidencesIMS with their respective
ResidenceCode values. Azores Islands and Portugal
geometries will be combined into one, as for the Madeira islands.
ResidencesIMS <- rbind(ResidencesIMS,
data.frame(Residence = c("British Indian Ocean Territory",
"Pitcairn Island",
"Norfolk Island",
"Christmas Island",
"Svalbard and Jan Mayen Islands",
"Cocos (Keeling) Islands"),
ResidenceCode = c(86, 612, 574, 162, 744, 166)))
GeoPRT <- WAB[WAB$Name %in% c("Portugal", "Azores Islands"), ]
GeoPRT <- st_union(GeoPRT)
PRT <- st_sf(geometry = GeoPRT, crs = ESRI_54030)
PRT <- PRT %>%
mutate(ISO3 = "PRT",
Status = "Member State",
ColorCode = "PRT",
Name = "Portugal",
Continent = "Europe",
Region = "Southern Europe",
ISO2 = "PT",
FName = "Portugal") %>%
select(-1, everything())
WAB <- WAB %>% filter(!Name %in% c("Portugal", "Azores Islands"))
WAB <- rbind(WAB, PRT)
RIMS_WAB <- left_join(WAB, ResidencesIMS, by = c("Name" = "Residence")) %>%
rename(Residence = Name) %>% select(4, 10, 1:3, 5:9)
RIMS_WAB_NoGeo <- RIMS_WAB %>% st_drop_geometry()
GGPLOT_RIMS_WAB <- ggplot() +
geom_sf(data = RIMS_WAB, aes(fill = ISO3)) +
labs(title = "World Administrative Boundaries (Source: OpenDataSoft)",
x = "Longitude", y = "Latitude") +
theme(legend.position = 'none') +
ggspatial::annotation_north_arrow(location = "tr", which_north = TRUE,
height = unit(1, "cm"), width = unit(1, "cm"),
style = north_arrow_fancy_orienteering(
line_col = NA,
fill = c("black", "black"),
text_size = 5))
GGPLOT_RIMS_WAB
There are 255 Residence | 247 ResidenceCode
| 247 ISO3 | 246 ISO2.
We have more Residence than in initial migrant stock
data which was 235.
datatable(RIMS_WAB_NoGeo[ , 1:3], options = list(scrollX = TRUE, autoWidth = TRUE),
caption = "Residences in Geo Data",
style = "bootstrap", rownames = FALSE) %>%
formatStyle(columns = 1, textAlign = "left") %>%
formatStyle(columns = c(2:3), textAlign = "center")We can make a simple interactive map using leaflet.
leaflet() %>%
addTiles() %>%
addPolygons(data = st_transform(RIMS_WAB, crs = '+proj=longlat +datum=WGS84'),
color = "blue", stroke = 1, opacity = 0.8, label = ~Residence)There is one little issue, we have to use WGS84.
3.3 Classification Data
United Nations Statistic Division (UNSD) data will be used for Regions Classification, SDG Classification and both Development Levels Classifications, while World Bank ones will be used for Income Levels Classification.
3.3.1 UNSD
UNSD <- read_excel("../Données/Regions/ONU/UNSD.xlsx") %>%
rename(SubRegionCode = 5, SubRegionName = 6, Area = 9, M49Code = 10, ISO2 = 11, ISO3 = 12) %>%
select(3:15)
UNSD <- UNSD[!is.na(UNSD$`Region Name`), ]There is a variable Region Name:
- Africa
- Americas
- Asia
- Europe
- Oceania
There are also variables SubRegionName and
Intermediate Region Name:
datatable(UNSD %>% group_by(`SubRegionName`) %>% summarise(Count = n()),
options = list(scrollX = TRUE, autoWidth = TRUE),
style = "bootstrap", rownames = FALSE) %>%
formatStyle(columns = 1, textAlign = "left") %>%
formatStyle(columns = 2, textAlign = "center")kable(UNSD %>% group_by(`Intermediate Region Name`) %>% summarise(Count = n()),
align = c("l", "c"))| Intermediate Region Name | Count |
|---|---|
| Caribbean | 28 |
| Central America | 8 |
| Eastern Africa | 22 |
| Middle Africa | 9 |
| South America | 16 |
| Southern Africa | 5 |
| Western Africa | 17 |
| NA | 142 |
There are 247 Area | 247 M49 Code | 247
ISO3.
There are also 3 variables corresponding to the 3 Development Levels:
- Least Developed Countries (LDCs)
- Land-Locked Developing Countries (LLDCs)
- Small Island Developing States (SIDS)
There is a small cross (x) if the area is part of one level (an area can belong to 2 levels at once).
We are going to combine RIMS_WAB according to ISO3 to
see if there are any matching problems.
UNSD$SubRegionName <- gsub("South-eastern Asia", "South-Eastern Asia", UNSD$SubRegionName)
RIMS_WAB_UNSD <- left_join(RIMS_WAB, UNSD, by = "ISO3")
RIMS_WAB_UNSD_NoGeo <- RIMS_WAB_UNSD %>% st_drop_geometry()
datatable(RIMS_WAB_UNSD_NoGeo[is.na(RIMS_WAB_UNSD_NoGeo$`Region Code`), ][ , c(1, 3, 16)],
options = list(scrollX = TRUE, autoWidth = TRUE,
columnDefs = list(list(targets = "_all", render = JS(
"function(data, type, row, meta) {",
" if (type === 'display' && data === null) {",
" return 'NA';",
" }",
" return data;",
"}")))),
style = "bootstrap", rownames = FALSE,
caption = "Non-Matching Rows from UNSD (R) in RIMS_WAB (L)") %>%
formatStyle(columns = 1, textAlign = "left") %>%
formatStyle(columns = c(2:3), textAlign = "center")UNSD_RIMS_WAB <- left_join(UNSD, RIMS_WAB, by = "ISO3")
UNSD_RIMS_WAB_NoGeo <- UNSD_RIMS_WAB %>% st_drop_geometry()
kable(UNSD_RIMS_WAB_NoGeo[is.na(UNSD_RIMS_WAB_NoGeo$Residence), ][, c(7, 10, 14)],
row.names = FALSE, align = c("l", "c", "c"),
caption = "Non-Matching Rows from RIMS_WAB (R) in UNSD (L)")| Area | ISO3 | Residence |
|---|---|---|
| Åland Islands | ALA | NA |
| Guernsey | GGY | NA |
| Isle of Man | IMN | NA |
| Jersey | JEY | NA |
The Åland Islands are included in Finland in RIMS_WAB. We can remove them from UNSD.
Isle of Man have two distinct ISO3: IMY (in RIMS_WAB) | IMN (in UNSD). The correct one is IMN.
Jersey and Guernsey have been grouped together as Channel Islands in RIMS_WAB, so we’re doing the same thing in UNSD.
For the others, we will need to replace missing values where possible.
We can also delete certain columns because same information is duplicated in another column:
- We have two
ISO2columns. Residence=Area.ResidenceCode=M49Code.Continent=Region Name.
Some changes are made in Continent and
Region for certain islands to match with UNSD
information.
UNSD <- UNSD %>% filter(!Area %in% c("Åland Islands", "Jersey"))
RIMS_WAB[which(RIMS_WAB$Residence == "Isle of Man"), "ISO3"] <- "IMN"
UNSD[which(UNSD$Area == "Guernsey"), "M49Code"] <- 830
UNSD[which(UNSD$Area == "Guernsey"), "ISO3"] <- "CHA"
UNSD[which(UNSD$Area == "Guernsey"), "ISO2"] <- NA
UNSD[which(UNSD$Area == "Guernsey"), "Area"] <- "Channel Islands"
RIMS_WAB_UNSD <- left_join(RIMS_WAB, UNSD, by = "ISO3")
RIMS_WAB_UNSD <- RIMS_WAB_UNSD %>%
select(-12, -17, -18, -19) %>%
rename(ISO2 = ISO2.x, ContinentCode = 11,
IntermediateRegionName = 15, IntermediateRegionCode = 14)
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "French Southern and Antarctic Territories"), "Continent"] <- "Africa"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "Christmas Island"), "Continent"] <- "Oceania"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "Christmas Island"), "Region"] <- "Australia and New Zealand"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "Bouvet Island"), "Continent"] <- "Americas"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "Heard Island and McDonald Islands"), "Continent"] <- "Oceania"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "Heard Island and McDonald Islands"), "Region"] <- "Australia and New Zealand"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "Cocos (Keeling) Islands"), "Continent"] <- "Oceania"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "Cocos (Keeling) Islands"), "Region"] <- "Australia and New Zealand"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "Cocos (Keeling) Islands"), "ContinentCode"] <- 9
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "South Georgia & the South Sandwich Islands"), "Continent"] <- "Americas"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "United States Minor Outlying Islands"), "Continent"] <- "Oceania"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "United States Minor Outlying Islands"), "Region"] <- "Micronesia"
RIMS_WAB_UNSD <- RIMS_WAB_UNSD %>%
select(1:3, 8, 6, 11, 7, 13, 12, 15, 14, 16:18, 4, 5, 9:10)
RIMS_WAB_UNSD$ContinentCode <- ifelse(is.na(RIMS_WAB_UNSD$ContinentCode),
ifelse(RIMS_WAB_UNSD$Continent == "Asia", 142,
ifelse(RIMS_WAB_UNSD$Continent == "Africa", 2,
ifelse(RIMS_WAB_UNSD$Continent == "Americas", 19,
RIMS_WAB_UNSD$ContinentCode))), RIMS_WAB_UNSD$ContinentCode)
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "French Southern and Antarctic Territories"), "Region"] <- "Eastern Africa"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "Bouvet Island"), "Region"] <- "South America"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "South Georgia & the South Sandwich Islands"), "Region"] <- "South America"
RIMS_WAB_UNSD$SubRegionName <- ifelse(is.na(RIMS_WAB_UNSD$SubRegionName),
ifelse(RIMS_WAB_UNSD$Region == "Eastern Asia", "Eastern Asia",
ifelse(RIMS_WAB_UNSD$Region == "Northern Africa", "Northern Africa",
ifelse(RIMS_WAB_UNSD$Region == "South-Eastern Asia", "South-Eastern Asia",
ifelse(RIMS_WAB_UNSD$Region == "Caribbean", "Latin America and the Caribbean",
RIMS_WAB_UNSD$SubRegionName)))), RIMS_WAB_UNSD$SubRegionName)
RIMS_WAB_UNSD$SubRegionCode <- ifelse(is.na(RIMS_WAB_UNSD$SubRegionCode),
ifelse(RIMS_WAB_UNSD$Region == "Eastern Asia", 30,
ifelse(RIMS_WAB_UNSD$Region == "Northern Africa", 15,
ifelse(RIMS_WAB_UNSD$Region == "South-Eastern Asia", 35,
ifelse(RIMS_WAB_UNSD$Region == "Caribbean", 419,
RIMS_WAB_UNSD$SubRegionCode)))), RIMS_WAB_UNSD$SubRegionCode)
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "Guantanamo"), "IntermediateRegionName"] <- "Caribbean"
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "Guantanamo"), "IntermediateRegionCode"] <- 29
RIMS_WAB_UNSD_NoGeo <- RIMS_WAB_UNSD %>% st_drop_geometry()3.3.2 World Bank
WorldBank <- read_excel("../Données/Regions/World Bank/WorldBank.xlsx",
sheet = "Country Analytical History", skip = 4)
WorldBank <- WorldBank[7:224, c(1, 2, 35)] %>% rename(ISO3 = 1, NameWB = 2, IncomeLevel = 3)There is a variable IncomeLevel corresponding to the 3/4
Income Levels:
- High-Income (H)
- Middle-Income
- Upper-Middle-Income (UM)
- Lower-Middle-Income (LM)
- Low-Income (L)
There are 218 NameWB | 218 ISO3.
RIMS_WAB_UNSD_WorldBank <- left_join(RIMS_WAB_UNSD, WorldBank, by = "ISO3")
RIMS_WAB_UNSD_WorldBank_NoGeo <- RIMS_WAB_UNSD_WorldBank %>% st_drop_geometry()
datatable(RIMS_WAB_UNSD_WorldBank_NoGeo[is.na(RIMS_WAB_UNSD_WorldBank_NoGeo$IncomeLevel), ][ , c(1, 3, 18:19)],
options = list(scrollX = TRUE, autoWidth = TRUE,
columnDefs = list(list(targets = "_all", render = JS(
"function(data, type, row, meta) {",
" if (type === 'display' && data === null) {",
" return 'NA';",
" }",
" return data;",
"}")))),
style = "bootstrap", rownames = FALSE,
caption = "Non-Matching Rows from World Bank (R) in RIMS_WAB_UNSD (L)") %>%
formatStyle(columns = 1, textAlign = "left") %>% # CHange tooo
formatStyle(columns = c(2, 3:4), textAlign = "center") # Change tooWorldBank_RIMS_WAB_UNSD <- left_join(WorldBank, RIMS_WAB_UNSD, by = "ISO3")
datatable(WorldBank_RIMS_WAB_UNSD[is.na(WorldBank_RIMS_WAB_UNSD$Residence), ][ , c(1:4)],
options = list(scrollX = TRUE, autoWidth = TRUE,
columnDefs = list(list(targets = "_all", render = JS(
"function(data, type, row, meta) {",
" if (type === 'display' && data === null) {",
" return 'NA';",
" }",
" return data;",
"}")))),
style = "bootstrap", rownames = FALSE,
caption = "Non-Matching Rows from RIMS_WAB_UNSD (R) in World Bank (L)") %>%
formatStyle(columns = 2, textAlign = "left") %>%
formatStyle(columns = c(1, 3:4), textAlign = "center")Channel Islands have two distinct ISO3: CHA (in
RIMS_WAB_UNSD) | CHI (in World Bank). There isn’t official ISO3 for
Channel Islands and GB-CHA was Channel Islands ISO2 until 2007. We
decided to keep CHI as ISO3 for Channel Islands (Abel and Cohen |
2019).
Kosovo declared independence from Serbia in 2008. Kosovo does is not
a UN member and has no official iso code. Kosovo is included in Serbia
in RIMS_WAB_UNSD. Kosovo geo is available in
rnaturalearth.
RIMS_WAB_UNSD[which(RIMS_WAB_UNSD$Residence == "Channel Islands"), "ISO3"] <- "CHI"
RIMS_WAB_UNSD_WorldBank <- left_join(RIMS_WAB_UNSD, WorldBank, by = "ISO3")
RIMS_WAB_UNSD_WorldBank_NoGeo <- RIMS_WAB_UNSD_WorldBank %>% st_drop_geometry()We can also delete NameWB because same information is
duplicated in Residence.
RIMS_WAB_UNSD_WorldBank <- RIMS_WAB_UNSD_WorldBank %>%
select(-19) %>% select(1:14, 19, 15:18)
RIMS_WAB_UNSD_WorldBank_NoGeo <- RIMS_WAB_UNSD_WorldBank %>% st_drop_geometry()3.4 Classification Graphics
RIMS_WAB_UNSD_WorldBank <- RIMS_WAB_UNSD_WorldBank %>%
mutate(ContinentalRegion = case_when(
Continent == "Africa" ~ "Africa",
Continent == "Asia" ~ "Asia",
Continent == "Europe" ~ "Europe",
Continent == "Oceania" ~ "Oceania",
Continent == "Americas" & SubRegionName == "Northern America" ~ "Northern America",
Continent == "Americas" & SubRegionName == "Latin America and the Caribbean" ~ "Latin America and the Caribbean"))
ContinentalRegions <- ggplot() +
geom_sf(data = RIMS_WAB_UNSD_WorldBank, aes(fill = ContinentalRegion)) +
labs(title = "6 Continental Regions (Source: OpenDataSoft | UNSD)",
x = "Longitude", y = "Latitude") +
ggspatial::annotation_north_arrow(location = "tr", which_north = TRUE,
height = unit(1, "cm"), width = unit(1, "cm"),
style = north_arrow_fancy_orienteering(
line_col = NA,
fill = c("black", "black"),
text_size = 5)) +
scale_fill_manual(values = c("#339900", "#FF9900", "#003366", "#FF0000", "#663399", "#FFCC00"),
labels = c("Africa", "Asia", "Europe", "Latin America and the Caribbean", "Northern America", "Oceania"),
name = "Continental Region")
ContinentalRegions
RIMS_WAB_UNSD_WorldBank <- RIMS_WAB_UNSD_WorldBank %>%
mutate(ClasseSDG = case_when(
Continent == "Europe" ~ "Europe and Northern America",
SubRegionName == "Northern America" ~ "Europe and Northern America",
SubRegionName == "Latin America and the Caribbean" ~ "Latin America and the Caribbean",
SubRegionName == "Australia and New Zealand" ~ "Australia and New Zealand",
SubRegionName == "Sub-Saharan Africa" ~ "Sub-Saharan Africa",
SubRegionName %in% c("Northern Africa", "Western Asia") ~ "Northern Africa and Western Asia",
SubRegionName %in% c("Central Asia", "Southern Asia") ~ "Central and Southern Asia",
SubRegionName %in% c("Eastern Asia", "South-Eastern Asia") ~ "Eastern and South-Eastern Asia",
SubRegionName %in% c("Melanesia", "Micronesia", "Polynesia") ~ "Oceania*"
))
GeoClasseSDG <- ggplot() +
geom_sf(data = RIMS_WAB_UNSD_WorldBank, aes(fill = ClasseSDG)) +
labs(title = "8 Geographic Regions based on Sustainable Development Goals (SDG) Classification (Source: OpenDataSoft | UN)",
x = "Longitude", y = "Latitude") +
ggspatial::annotation_north_arrow(location = "tr", which_north = TRUE,
height = unit(1, "cm"), width = unit(1, "cm"),
style = north_arrow_fancy_orienteering(
line_col = NA,
fill = c("black", "black"),
text_size = 5)) +
scale_fill_manual(values = c("#FF0000", "#FF6600", "#339900", "#003366", "#33CCFF", "#FF9900", "#CC0000", "#FF3399"),
name = "SDG Classification")
GeoClasseSDG
RIMS_WAB_UNSD_WorldBank <- RIMS_WAB_UNSD_WorldBank %>%
mutate(MoreLess = case_when(
Continent == "Europe" ~ "More Developed",
SubRegionName == "Northern America" ~ "More Developed",
Continent == "Asia" & Residence == "Japan" ~ "More Developed",
Continent == "Oceania" & Residence %in% c("Australia", "New Zealand") ~ "More Developed",
Continent == "Africa" ~ "Less Developed",
Continent == "Asia" & !(Residence == "Japan") ~ "Less Developed",
SubRegionName == "Latin America and the Caribbean" ~ "Less Developed",
Continent == "Oceania" & !(Residence %in% c("Australia", "New Zealand")) ~ "Less Developed"))
MoreLessDevt <- ggplot() +
geom_sf(data = RIMS_WAB_UNSD_WorldBank, aes(fill = MoreLess)) +
labs(title = "2 Development Levels (Source: OpenDataSoft)",
x = "Longitude", y = "Latitude") +
ggspatial::annotation_north_arrow(location = "tr", which_north = TRUE,
height = unit(1, "cm"), width = unit(1, "cm"),
style = north_arrow_fancy_orienteering(
line_col = NA,
fill = c("black", "black"),
text_size = 5)) +
scale_fill_manual(values = c("#FFCC00", "#339900"),
labels = c("Less Developed", "More Developed"),
name = "Development Level")
MoreLessDevt
Outdated classification…
RIMS_WAB_UNSD_WorldBank <- RIMS_WAB_UNSD_WorldBank %>%
mutate(
DevtLevel = case_when(
`Least Developed Countries (LDC)` == 'x' &
`Land Locked Developing Countries (LLDC)` == 'x' &
`Small Island Developing States (SIDS)` == 'x' ~ 'LDC | LLDC | SIDS',
`Least Developed Countries (LDC)` == 'x' &
`Land Locked Developing Countries (LLDC)` == 'x' &
is.na(`Small Island Developing States (SIDS)`) ~ 'LDC | LLDC',
`Least Developed Countries (LDC)` == 'x' &
is.na(`Land Locked Developing Countries (LLDC)`) &
`Small Island Developing States (SIDS)` == 'x' ~ 'LDC | SIDS',
is.na(`Least Developed Countries (LDC)`) &
`Land Locked Developing Countries (LLDC)` == 'x' &
`Small Island Developing States (SIDS)` == 'x' ~ 'LLDC | SIDS',
`Least Developed Countries (LDC)` == 'x' &
is.na(`Land Locked Developing Countries (LLDC)`) &
is.na(`Small Island Developing States (SIDS)`) ~ 'LDC',
is.na(`Least Developed Countries (LDC)`) &
`Land Locked Developing Countries (LLDC)` == 'x' &
is.na(`Small Island Developing States (SIDS)`) ~ 'LLDC',
is.na(`Least Developed Countries (LDC)`) &
is.na(`Land Locked Developing Countries (LLDC)`) &
`Small Island Developing States (SIDS)` == 'x' ~ 'SIDS',
)
)
DevtLevels <- ggplot() +
geom_sf(data = RIMS_WAB_UNSD_WorldBank, aes(fill = DevtLevel)) +
labs(title = "3/5 Development Levels (Source: OpenDataSoft | UNSD)",
x = "Longitude", y = "Latitude") +
ggspatial::annotation_north_arrow(location = "tr", which_north = TRUE,
height = unit(1, "cm"), width = unit(1, "cm"),
style = north_arrow_fancy_orienteering(
line_col = NA,
fill = c("black", "black"),
text_size = 5)) +
scale_fill_manual(values = c("#FF0000", "#FFCC00", "#CC99FF", "#339900", "#33CCFF", "#999999"),
name = "Development Level")
DevtLevels
Make one map for each (LDC | LLDC | SIDS)…
IncomeLevels <- ggplot() +
geom_sf(data = RIMS_WAB_UNSD_WorldBank, aes(fill = IncomeLevel)) +
labs(title = "3/4 Income Levels based on Gross National Income (Source: OpenDataSoft | World Bank 2020)",
x = "Longitude", y = "Latitude") +
ggspatial::annotation_north_arrow(location = "tr", which_north = TRUE,
height = unit(1, "cm"), width = unit(1, "cm"),
style = north_arrow_fancy_orienteering(
line_col = NA,
fill = c("black", "black"),
text_size = 5)) +
scale_fill_manual(values = c("#339900", "#663399", "#CC99FF", "#99FF66", "#999999"),
labels = c("H", "L", "LM", "UM", "NA"),
name = "Income Level")
IncomeLevels
Make them interactive…
4 Graphics
Visualization at regional level will be done to begin with using SDG Classification.
We need to sum IMS from these 8 regions (IMSTSUM) to
compute share of IMS.
SDG_Classification = c(947, 1833, 921, 1832, 1830, 1835, 927, 1829)
imsflt <- ims %>% filter(Residence == "WORLD", BirthCode %in% SDG_Classification,
Year == c(1990, 1995, 2000, 2005, 2010, 2015, 2020)) %>%
distinct(BirthCode, Year, .keep_all = TRUE)
imss <- imsflt %>%
group_by(Year) %>%
summarize(IMSTSUM = sum(IMST), IMSMSUM = sum(IMSM), IMSFSUM = sum(IMSF))
imssww <- ims %>% filter(Residence == "WORLD", Birth == "WORLD",
Year == c(1990, 1995, 2000, 2005, 2010, 2015, 2020))
imss <- merge(imss, imssww, by = "Year") %>%
rename(IMSTWW = IMST, IMSMWW = IMSM, IMSFWW = IMSF) %>%
mutate(MISSINGT_SHARE = round(((IMSTWW - IMSTSUM) / IMSTWW) * 100, 3),
MISSINGM_SHARE = round(((IMSMWW - IMSMSUM) / IMSMWW) * 100, 3),
MISSINGF_SHARE = round(((IMSFWW - IMSFSUM) / IMSFWW) * 100, 3)) %>%
select(5, 6, 7, 8, 1, 9, 2, 12, 10, 3, 13, 11, 4, 14)
datatable(imss[, c(1, 3, 5:14)], options = list(scrollX = TRUE, autoWidth = TRUE),
style = "bootstrap", rownames = FALSE,
caption = "World's Total Migration Stocks (1990-2020)") %>%
formatStyle(columns = c(1, 2), textAlign = "left") %>%
formatStyle(columns = c(3:14), textAlign = "center")IMSTSUM is lower than the World-World IMS
(IMSTWW). There is 4-5.5% missing data
(MISSINGT_SHARE) in IMSTSUM. Where are these
4-5.5% ? Xlsx file include a row with IMS born in
Other for each country but not at regional level, so we are
going to sum all these Other rows (IMSTOSUM)
to see if the number matches with MISSINGT_SHARE.
imso <- ims %>% filter(Birth == "Other")
imsos <- imso %>%
group_by(Year) %>%
summarize(IMSTOSUM = sum(IMST), IMSMOSUM = sum(IMSM), IMSFOSUM = sum(IMSF))
imsos$Residence <- imss$Residence
imsos$ResidenceCode <- imss$ResidenceCode
imsos$Birth <- "Other"
imsos$BirthCode <- 2003
imsos <- imsos %>% select(5:8, everything())
kable(imsos[ , c(1, 3, 5:8)], row.names = FALSE,
align = c("l", "l", "c", "c", "c", "c"),
caption = "World's Migration Stocks born in Other (1990-2020)")| Residence | Birth | Year | IMSTOSUM | IMSMOSUM | IMSFOSUM |
|---|---|---|---|---|---|
| WORLD | Other | 1990 | 8640334 | 4408990 | 4231344 |
| WORLD | Other | 1995 | 7477497 | 3776224 | 3701273 |
| WORLD | Other | 2000 | 7063730 | 3593958 | 3469772 |
| WORLD | Other | 2005 | 7820665 | 4096864 | 3723801 |
| WORLD | Other | 2010 | 8757236 | 4601492 | 4155744 |
| WORLD | Other | 2015 | 10918842 | 5685512 | 5233330 |
| WORLD | Other | 2020 | 12657151 | 6533198 | 6123953 |
imsos <- imsos %>% rename(IMST = IMSTOSUM, IMSM = IMSMOSUM, IMSF = IMSFOSUM)
imsflt <- rbind(imsflt, imsos)IMSTWW = IMSTSUM + IMSTOSUM.
We can add a ninth region called Other to make our
visualization. Now, we are able to compute the share of IMS for each
region (IMST_SHARE).
share_values <- list()
for (row in 1:nrow(imsflt)) {
YYYY <- imsflt$Year[row]
IMSS <- imss$IMSTWW[imss$Year == YYYY]
SHARE <- (imsflt$IMST[row]/IMSS) * 100
share_values[[row]] <- SHARE
}
imsflt$IMST_SHARE <- unlist(share_values)
kable(imsflt[50:56, c(1, 3, 5:9)], row.names = FALSE,
align = c("l", "l", "c", "c", "c", "c", "c"),
caption = "World's Migration Stocks (%) born in Europe and Northern America (1990-2020)",
digits = c(0, 0, 0, 0, 0, 0, 3))| Residence | Birth | Year | IMST | IMSM | IMSF | IMST_SHARE |
|---|---|---|---|---|---|---|
| WORLD | Europe and Northern America | 1990 | 50532413 | 23704534 | 26827879 | 33.031 |
| WORLD | Europe and Northern America | 1995 | 51582164 | 24148758 | 27433406 | 31.981 |
| WORLD | Europe and Northern America | 2000 | 52877703 | 24676104 | 28201599 | 30.524 |
| WORLD | Europe and Northern America | 2005 | 54849090 | 25682798 | 29166292 | 28.650 |
| WORLD | Europe and Northern America | 2010 | 58681104 | 27337252 | 31343852 | 26.555 |
| WORLD | Europe and Northern America | 2015 | 60592402 | 28229706 | 32362696 | 24.436 |
| WORLD | Europe and Northern America | 2020 | 67601621 | 31770385 | 35831236 | 24.092 |
The share of IMS in the World born in Europe and NA is decreasing, from 33% in 1990 to 24% in 2020.
4.1 Visualization
To visualize IMST_SHARE we will use 3 libraries
(ggplot2 | plotly | ggiraph). A
stacked barplot is used to represent the share of international migrant
stocks in a residence area from \(n\)
birth area, while a line plot will show the evolution of IMS from 1990
to 2020.
4.1.1 GGPLOT2
ggplot2 make nice static graphics easily.
SDG_Colors <- c("#FF0000", "#FF6600", "#339900", "#003366", "#33CCFF", "#FF9900", "#CC0000", "#999999", "#FF3399")
GGPLOT_SBAR <- ggplot(imsflt, aes(x = Year, y = IMST_SHARE, fill = Birth,
text = paste(" Residence:", Residence, "<br>",
"Birth:", Birth, "<br>",
"Year:", Year, "<br>",
"International Migrant Stock:", round(IMST_SHARE, 3), "%"))) +
geom_bar(stat = "identity") +
labs(title = "Share of International Migrant Stocks in the World from 8 Areas (1990-2020)",
y = "International Migrant Stock (%)") +
scale_fill_manual(values = SDG_Colors) +
theme_minimal() +
theme(legend.position = "right")
GGPLOT_SBAR
This plot allow us to see fastly the evolution of share from 1990 to 2020, it’s falling for Europe and NA (see previous section), increasing for Latin America and the Caribbean and it’s quite constant for Central and Southern Asia or Sub-Saharan Africa. It’s also a good way to see from which area IMS most often originate.
Birth regions are displayed in A-Z order. If we want to reorder them, we will have to use factor levels.
There is some negative points to this graphics. We can’t see correctly Australia and New Zealand or Oceania*. We don’t really know accurately how many % each color represent in a bar because this is a static plot, so an interactive one may be necessary.
The shares are only a part of the information, we need to know the stocks.
GGPLOT_LINE <- ggplot(imss, aes(x = Year)) +
geom_line(aes(y = IMSTWW, color = "IMSTWW", group = "IMSTWW")) +
geom_line(aes(y = IMSMWW, color = "IMSMWW", group = "IMSMWW")) +
geom_line(aes(y = IMSFWW, color = "IMSFWW", group = "IMSFWW")) +
geom_point(aes(y = IMSTWW, color = "IMSTWW", group = "IMSTWW")) +
geom_point(aes(y = IMSMWW, color = "IMSMWW", group = "IMSMWW")) +
geom_point(aes(y = IMSFWW, color = "IMSFWW", group = "IMSFWW")) +
labs(title = "Evolution of International Migrant Stocks in the World (1990-2020)",
y = "International Migrant Stock") +
scale_y_continuous(labels = scales::label_number(scale = 1e-6, suffix = "M")) +
theme_minimal() +
scale_color_manual(values = c("IMSTWW" = "black", "IMSMWW" = "blue", "IMSFWW" = "pink"),
labels = c("IMSTWW" = "Both Sexes Combined", "IMSMWW" = "Males", "IMSFWW" = "Females")) +
theme(legend.position = "top", legend.title = element_blank())
GGPLOT_LINE
We can clearly follow the evolution of IMS thanks to this line plot, an evolution which is stronger since 2005. Stocks are fairly balanced between males and females (1990-2005), then the gap widens in favor of males (2005-2020).
Problem is the same as in the stacked barplot, the lack of interactivity to get the real number at each observation. We also have to rescale the number manually to add the “M” suffix to it.
grid.arrange(GGPLOT_SBAR, GGPLOT_LINE, ncol=1, nrow=2, heights = c(1, 0.8))
It’s easy to combine the stacked barplot and the line plot together
thanks to grid.arrange(). One positive point is the
possibility of having two distinct legends which is not always the
case.
Maybe it’s not really useful because we might prefer to have two unique plots in the shiny app. It offers more flexibility.
4.1.2 PLOTLY
plotly enable us to make interactive plots after a
simple data manipulation.
imsfltpltoly <- imsflt %>%
group_by(Year, Residence, Birth) %>%
summarise(IMST_SHARE = sum(IMST_SHARE), IMST = sum(IMST))
kable(imsfltpltoly[1:9, 1:4], row.names = FALSE,
align = c("l", "l", "l", "c"),
caption = "Share of International Migrant Stocks in the World from 8 Area in 1990",
digits = c(0, 0, 0, 3))| Year | Residence | Birth | IMST_SHARE |
|---|---|---|---|
| 1990 | WORLD | Australia and New Zealand | 0.450 |
| 1990 | WORLD | Central and Southern Asia | 20.331 |
| 1990 | WORLD | Eastern and South-Eastern Asia | 9.657 |
| 1990 | WORLD | Europe and Northern America | 33.031 |
| 1990 | WORLD | Latin America and the Caribbean | 9.984 |
| 1990 | WORLD | Northern Africa and Western Asia | 10.697 |
| 1990 | WORLD | Oceania (excluding Australia and New Zealand) | 0.178 |
| 1990 | WORLD | Other | 5.648 |
| 1990 | WORLD | Sub-Saharan Africa | 10.024 |
Here we need to have data for a given year and all
Residence-Birth pairs in a row which is the
contrary of the data used with ggplot where we have data
for a given Residence-Birth pair and all years
in a row.
PLOTLY_SBAR <- plot_ly(data = imsfltpltoly, x = ~Year, y = ~IMST_SHARE,
color = ~Birth, type = "bar", colors = SDG_Colors, legendgroup = 'Birth',
text = with(imsfltpltoly, paste(" Residence:", Residence, "<br>",
"Birth:", Birth, "<br>",
"Year:", Year, "<br>",
"International Migrant Stock:", round(IMST_SHARE, 3), "%")), hoverinfo = "text") %>%
layout(barmode = "stack",
yaxis = list(title = "International Migrant Stock (%)"),
title = "Share of International Migrant Stocks in the World from 8 Areas (1990-2020)",
legend = list(title = list(text='<b> Birth </b>'), x=1, y=0.5))
PLOTLY_SBARThere is a download button which is something we want to save the plot as png in the shiny app. We can have all the information we want by simply moving the mouse on the plot: Residence | Birth | Year | %. But the information are displayed in a vector so we need to manually change the hover text.
Birth regions are displayed in Z-A order, which is the opposite of
ggplot.
PLOTLY_LINE <- plot_ly(data = imss, x = ~Year, y=~IMSTWW, type = 'scatter', mode = 'lines+markers', name = 'Both Sexes Combined', line = list(color = 'black'), marker = list(color = 'black'), text = with(imss, paste(" Residence:", Residence, "<br>",
"Birth:", Birth, "<br>",
"Year:", Year, "<br>",
"International Migrant Stock (Both Sexes Combined):", paste0(format(IMSTWW, big.mark = " ", scientific = FALSE), "M"))), hoverinfo = "text") %>%
add_trace(x = ~Year, y = ~IMSMWW, name = 'Males', line = list(color = 'blue'), marker = list(color = 'blue'), text = with(imss, paste(" Residence:", Residence, "<br>",
"Birth:", Birth, "<br>",
"Year:", Year, "<br>",
"International Migrant Stock (Males):", paste0(format(IMSMWW, big.mark = " ", scientific = FALSE), "M"))), hoverinfo = "text") %>%
add_trace(x = ~Year, y = ~IMSFWW, name = 'Females', line = list(color = 'pink'), marker = list(color = 'pink'), text = with(imss, paste(" Residence:", Residence, "<br>",
"Birth:", Birth, "<br>",
"Year:", Year, "<br>",
"International Migrant Stock (Females):", paste0(format(IMSFWW, big.mark = " ", scientific = FALSE), "M"))), hoverinfo = "text") %>%
layout(title = "Evolution of International Migrant Stocks in the World (1990-2020)",
xaxis = list(title = "Year"),
yaxis = list(title = "International Migrant Stock"),
legend = list(x = 0.5, y = 1.05, xanchor = "center", orientation = "h"))
PLOTLY_LINEWe don’t need to rescale numbers to get “M” suffix on the y-axis but
we need to do it for the hover text. Furthermore, the legend isn’t
movable as in ggplot2.
PLOTLY_SBAR <- PLOTLY_SBAR %>%
layout(annotations = list(x = 0.5, y = 1.05, xref = "paper", yref = "paper", text = "Share of International Migrant Stocks in the World from 8 Areas (1990-2020)", showarrow = FALSE))
PLOTLY_LINE <- PLOTLY_LINE %>%
layout(title = "",
annotations = list(x = 0.5, y = 1, xref = "paper", yref = "paper", text = "Evolution of International Migrant Stocks in the World (1990-2020)", showarrow = FALSE),
legend = list(title = list(text='<b> </b>'),
x = 1.05, xanchor = "left", y = 0.5, yanchor = "middle", orientation = "v"))
PLOTLY <- subplot(PLOTLY_SBAR, PLOTLY_LINE, nrows = 2, heights = c(0.65, 0.35),
titleX = TRUE, titleY = TRUE, shareX = TRUE) %>%
layout(showlegend = TRUE)
PLOTLYIt’s also easy to combine the stacked barplot and the line plot
together thanks to subplot(). One positive point is the
possibility of having a share x-axis easily.
There is some negative points, we can’t have two distinct legends in
plotly, we also have to force to have the y-axis title for
both plots. It automatically keep legend position and title of the last
plot as the title of the whole subplot, so we have to manage it.
4.1.3 GGPLOTLY
We can also transform a ggplot2 graphics into a
plotly one using ggplotly()
GGPLOTLY_SBAR <- ggplotly(GGPLOT_SBAR)
GGPLOTLY_SBARWe just have to change the hover text inside the initial
ggplot2 plot like in plotly one. There is an
important problem is that we can’t remove the default hover text.
GGPLOTLY_LINE <- ggplotly(GGPLOT_LINE) %>%
layout(legend = list(title = "", x = 0.5, y = 1.05, xanchor = "center", orientation = "h"))
for (i in 1:length(GGPLOTLY_LINE$x$data)) {
GGPLOTLY_LINE$x$data[[i]]$name <- c("Both Sexes Combined", "Males", "Females")[i]
}
GGPLOTLY_LINEThe custom ggplot2 legend which is not support by
plotly so we have to remake one using plotly
code. The text argument is not support in
aes() inside geom_point() so we can’t modify
the hover text.
GGPLOTLY_SBAR <- GGPLOTLY_SBAR %>%
layout(annotations = list(x = 0.5, y = 1.05, xref = "paper", yref = "paper", text = "Share of International Migrant Stocks in the World from 8 Areas (1990-2020)", showarrow = FALSE))
GGPLOTLY_LINE <- GGPLOTLY_LINE %>%
layout(title = "",
annotations = list(x = 0.5, y = 1, xref = "paper", yref = "paper", text = "Evolution of International Migrant Stocks in the World (1990-2020)", showarrow = FALSE),
legend = list(title = list(text='<b> </b>'),
x = 1.05, xanchor = "left", y = 0.5, yanchor = "middle", orientation = "v"))
GGPLOTLY <- subplot(GGPLOTLY_SBAR, GGPLOTLY_LINE, nrows = 2, heights = c(0.65, 0.35),
titleX = TRUE, titleY = TRUE, shareX = TRUE) %>%
layout(showlegend = TRUE)
GGPLOTLYSame things as with plotly.
4.1.4 GGIRAPH
ggiraph use ggplot2 and we just have to use
geom_bar_interactive instead of geom_bar to
make an interactive plot, as example, otherwise it’s the same
syntax.
GGIRAPH_SBAR <- ggplot(imsflt) +
geom_bar_interactive(aes(x = Year, y = IMST_SHARE, fill = Birth,
tooltip = paste("Residence:", Residence,
"<br>Birth:", Birth,
"<br>Year:", Year,
"<br>International Migrant Stock:", round(IMST_SHARE, 3), "%"))
, stat = "identity") +
labs(title = "Share of International Migrant Stocks in the World from 8 Areas (1990-2020)",
y = "International Migrant Stock (%)") +
scale_fill_manual(values = SDG_Colors) +
theme_minimal() +
theme(legend.position = "right")
girafe(ggobj = GGIRAPH_SBAR)We use tooltip argument to modify the hover text as we
need easily (there is no background color like in plotly).
We also have automatically a download button.
GGIRAPH_LINE <- ggplot(imss) +
geom_line_interactive(aes(x = Year, y = IMSTWW, color = "IMSTWW", group = "IMSTWW")) +
geom_line_interactive(aes(x = Year, y = IMSMWW, color = "IMSMWW", group = "IMSMWW")) +
geom_line_interactive(aes(x = Year, y = IMSFWW, color = "IMSFWW", group = "IMSFWW")) +
geom_point_interactive(aes(x = Year, y = IMSTWW, color = "IMSTWW", group = "IMSTWW",
tooltip = paste("Residence:", Residence,
"<br>Birth:", Birth,
"<br>Year:", Year,
"<br>International Migrant Stock (Both Sexes Combined):", paste0(format(IMSTWW, big.mark = " ", scientific = FALSE), "M")))) +
geom_point_interactive(aes(x = Year, y = IMSMWW, color = "IMSMWW", group = "IMSMWW",
tooltip = paste("Residence:", Residence,
"<br>Birth:", Birth,
"<br>Year:", Year,
"<br>International Migrant Stock (Males):", paste0(format(IMSMWW, big.mark = " ", scientific = FALSE), "M")))) +
geom_point_interactive(aes(x = Year, y = IMSFWW, color = "IMSFWW", group = "IMSFWW",
tooltip = paste("Residence:", Residence,
"<br>Birth:", Birth,
"<br>Year:", Year,
"<br>International Migrant Stock (Females):", paste0(format(IMSFWW, big.mark = " ", scientific = FALSE), "M")))) +
labs(title = "Evolution of International Migrant Stocks in the World (1990-2020)",
y = "International Migrant Stock") +
scale_y_continuous(labels = scales::label_number(scale = 1e-6, suffix = "M")) +
theme_minimal() +
scale_color_manual(values = c("IMSTWW" = "black", "IMSMWW" = "blue", "IMSFWW" = "pink"),
labels = c("IMSTWW" = "Both Sexes Combined", "IMSMWW" = "Males", "IMSFWW" = "Females")) +
theme(legend.position = "top", legend.title = element_blank())
girafe(ggobj = GGIRAPH_LINE)It’s not possible to combine two ggiraph plots
together.
4.1.5 Resume
We might use plotly or ggplot2 combine with
ggiraph to make our stacked barplots and line plots.
LibrariesComparison <- data.frame(
LIBRARY = c("GGPLOT2", "PLOTLY", "GGPLOTLY", "GGIRAPH"),
POSITIVE = c("Simple Syntax | Nice Graphics", "Interactive | DownLoad Button | Hover Text | Share X-axis", "Interactive | DownLoad Button", "Interactive | DownLoad Button | Hover Text | GGPLOT2 Syntax"),
NEGATIVE = c("Static", "Legend Issues", "Legend Issues | Hover Text", "-"))
datatable(LibrariesComparison, options = list(scrollX = TRUE, autoWidth = TRUE),
style = "bootstrap", rownames = FALSE,
caption = "Graphics' Libraries Comparison")4.2 Stacked Barplots
4.2.1 IMS in Million
Line Plot is not the only way to represent the evolution of IMS. We can aslo make a stacked barplot in Million instead of in %.
4.2.1.1 PLOTLY
PLOTLY_SBARM <- plot_ly(data = imsfltpltoly, x = ~Year, y = ~IMST,
color = ~Birth, type = "bar", colors = SDG_Colors, legendgroup = 'Birth',
text = with(imsfltpltoly, paste(" Residence:", Residence, "<br>",
"Birth:", Birth, "<br>",
"Year:", Year, "<br>",
"International Migrant Stock:", paste0(format(IMST, big.mark = " ", scientific = FALSE), "M"))), hoverinfo = "text") %>%
layout(barmode = "stack",
yaxis = list(title = "International Migrant Stock"),
title = "Evolution of International Migrant Stocks in the World from 8 Areas (1990-2020)",
legend = list(title = list(text='<b> Birth </b>'), x=1, y=0.5))
PLOTLY_SBARM4.2.1.2 GGIRAPH
GGIRAPH_SBARM <- ggplot(imsflt) +
geom_bar_interactive(aes(x = Year, y = IMST, fill = Birth,
tooltip = paste("Residence:", Residence,
"<br>Birth:", Birth,
"<br>Year:", Year,
"<br>International Migrant Stock:",
paste0(format(IMST, big.mark = " ", scientific = FALSE), "M"))),
position = "stack", stat = "identity") +
labs(title = "Evolution of International Migrant Stocks in the World from 8 Areas (1990-2020)",
y = "International Migrant Stock") +
scale_fill_manual(values = SDG_Colors) +
scale_y_continuous(labels = scales::label_number(scale = 1e-6, suffix = "M")) +
theme_minimal() +
theme(legend.position = "right")
girafe(ggobj = GGIRAPH_SBARM)5 Application
Version 1.0 :
- Some data manipulations take time (e.g. calculation of migrant shares | …)
- Some data manipulations are unnecessary in ui/server (e.g. | …)
Need to do it to avoid long update/load times on client/server side.
5.1 IMS.DATA
- Income Levels with uppercase in
ResidenceandBirth - Calculation of variables to be used in the application:
- Share of international migrant stock
IMS.DATA <- ims %>%
mutate(
Residence = case_when(
Residence == "High-income countries" ~ "High-Income Countries",
Residence == "Upper-middle-income countries" ~ "Upper-Middle-Income Countries",
Residence == "Lower-middle-income countries" ~ "Lower-Middle-Income Countries",
Residence == "Low-income countries" ~ "Low-Income Countries",
TRUE ~ Residence),
Birth = case_when(
Birth == "High-income countries" ~ "High-Income Countries",
Birth == "Upper-middle-income countries" ~ "Upper-Middle-Income Countries",
Birth == "Lower-middle-income countries" ~ "Lower-Middle-Income Countries",
Birth == "Low-income countries" ~ "Low-Income Countries",
TRUE ~ Birth)
)
IMS.DATA.SHARE <- data.frame(
Residence = character(0),
ResidenceCode = numeric(0),
Birth = character(0),
BirthCode = numeric(0),
Year = character(0),
IMST = numeric(0),
IMSM = numeric(0),
IMSF = numeric(0),
IMST_SHARE = numeric(0),
IMSM_SHARE = numeric(0),
IMSF_SHARE = numeric(0),
stringsAsFactors = FALSE
)
SHAREs <- function(DATA, FDATA, VARIABLE) {
SValues <- list()
for (ROw in 1:nrow(FDATA)) {
YYYY <- FDATA$Year[ROw]
IMSS <- unique(DATA[DATA$ResidenceCode == SelectedRCode &
DATA$BirthCode == 900 & DATA$Year == YYYY, VARIABLE])
SHARE <- (FDATA[[VARIABLE]][ROw]/IMSS) * 100
SValues[[ROw]] <- SHARE
}
return(SValues)
}
OTHERs <- function(DATA, iMSSC, YEAR, VARIABLE, SVARIABLE, NROW) {
IMSS <- DATA[DATA$ResidenceCode == SelectedRCode &
DATA$BirthCode == 900 & DATA$Year == YEAR, VARIABLE]
IMS.OTHER <- IMSS[[1]] - iMSSC[[SVARIABLE]][NROW]
return(IMS.OTHER)
}
# for (ROW in 1:nrow(IMS.DATA)) {
# SelectedRCode <- IMS.DATA$ResidenceCode[ROW]
# if (SelectedRCode %in% IMS.DATA.SHARE$ResidenceCode) {
# # print("Done!")
# } else {
# IMS.FDATA <- IMS.DATA %>%
# filter(ResidenceCode == SelectedRCode,
# Year == c(1990, 1995, 2000, 2005, 2010, 2015, 2020)) %>%
# distinct(BirthCode, Year, .keep_all = TRUE)
# if (SelectedRCode < 900) {
# print(paste("PROGRESSION:", ROW, "/", nrow(IMS.DATA)))
#
# SValues.IMST <- SHAREs(DATA = IMS.DATA, FDATA = IMS.FDATA, VARIABLE = "IMST")
# IMS.FDATA$IMST_SHARE <- unlist(SValues.IMST)
#
# SValues.IMSM <- SHAREs(DATA = IMS.DATA, FDATA = IMS.FDATA, VARIABLE = "IMSM")
# IMS.FDATA$IMSM_SHARE <- unlist(SValues.IMSM)
#
# SValues.IMSF <- SHAREs(DATA = IMS.DATA, FDATA = IMS.FDATA, VARIABLE = "IMSF")
# IMS.FDATA$IMSF_SHARE <- unlist(SValues.IMSF)
#
# IMS.DATA.SHARE <- rbind(IMS.DATA.SHARE, IMS.FDATA)
# } else {
# print(paste("PROGRESSION:", ROW, "/", nrow(IMS.DATA)))
#
# fDATA <- IMS.FDATA %>%
# filter(BirthCode < 900,
# Year == c(1990, 1995, 2000, 2005, 2010, 2015, 2020)) %>%
# distinct(BirthCode, Year, .keep_all = TRUE)
#
# IMSSC <- fDATA %>%
# group_by(Year) %>%
# summarize(IMSTSUM = sum(IMST), IMSMSUM = sum (IMSM), IMSFSUM = sum(IMSF))
# for (Row in 1:nrow(IMSSC)) {
# YYYY <- IMSSC$Year[Row]
#
# IMST.OTHER <- OTHERs(DATA = IMS.DATA, iMSSC = IMSSC, YEAR = YYYY,
# VARIABLE = "IMST", SVARIABLE = "IMSTSUM", NROW = Row)
# IMSM.OTHER <- OTHERs(DATA = IMS.DATA, iMSSC = IMSSC, YEAR = YYYY,
# VARIABLE = "IMSM",SVARIABLE = "IMSMSUM", NROW = Row)
# IMSF.OTHER <- OTHERs(DATA = IMS.DATA, iMSSC = IMSSC, YEAR = YYYY,
# VARIABLE = "IMSF",SVARIABLE = "IMSFSUM", NROW = Row)
#
# IMS.FDATA <- rbind(IMS.FDATA, list(unique(IMS.FDATA$Residence),
# SelectedRCode, "Other", 2003,
# YYYY, IMST.OTHER, IMSM.OTHER, IMSF.OTHER))
# }
# SValues.IMST <- SHAREs(DATA = IMS.DATA, FDATA = IMS.FDATA, VARIABLE = "IMST")
# IMS.FDATA$IMST_SHARE <- unlist(SValues.IMST)
#
# SValues.IMSM <- SHAREs(DATA = IMS.DATA, FDATA = IMS.FDATA, VARIABLE = "IMSM")
# IMS.FDATA$IMSM_SHARE <- unlist(SValues.IMSM)
#
# SValues.IMSF <- SHAREs(DATA = IMS.DATA, FDATA = IMS.FDATA, VARIABLE = "IMSF")
# IMS.FDATA$IMSF_SHARE <- unlist(SValues.IMSF)
#
# IMS.DATA.SHARE <- rbind(IMS.DATA.SHARE, IMS.FDATA)
# }
# }
# }5.2 GeoDATA
RegionbecomesSIRegioninGeoDATASIRegionCodebased onSubRegionCodeandIntermediateRegionCodeinGeoDATA- Combine countries/territories geometries into areas :
- Continental Regions
- Continental Sub-Regions
- Geographic Regions (SDG)
- Income Levels
Label|LabelRwill be hover text used onleafletmaps
We assign a color code (CLRCode) to regions :
- Continental Sub-Regions colors are based on Continental Regions ones
- Geographic Regions (SDG) colors come from UN
- Income Levels colors come from World Bank
Factor levels will be set in DATA.R because not saved
when st_write().
We also assign CLRCode to countries based on Continental
Regions colors, except for territories without a
ResidenceCode.
GeoDATA <- RIMS_WAB_UNSD_WorldBank %>%
rename(SIRegion = Region) %>%
mutate(
SIRegionCode = ifelse(is.na(IntermediateRegionCode), SubRegionCode, IntermediateRegionCode),
Label = paste0("<b>Name: </b>", Residence, "<br>",
"<b>ISO Num.: </b>", ResidenceCode, "<br>",
"<b>ISO2: </b>", ISO2, "<br>",
"<b>ISO3: </b>", ISO3)) %>%
select(1:2, 4, 3, 5:7, 24, 12:15, 20:23, 25, 19)
GeoRDATA.CRs <- GeoDATA %>%
group_by(ContinentalRegion) %>%
summarize(
Continent = Mode(Continent),
ContinentCode = Mode(ContinentCode),
geometry = st_combine(geometry)) %>%
mutate(
Residence = case_when(
ContinentalRegion == "Africa" ~ "AFRICA",
ContinentalRegion == "Asia" ~ "ASIA",
ContinentalRegion == "Europe" ~ "EUROPE",
ContinentalRegion == "Latin America and the Caribbean" ~ "LATIN AMERICA AND THE CARIBBEAN",
ContinentalRegion == "Northern America" ~ "NORTHERN AMERICA",
ContinentalRegion == "Oceania" ~ "OCEANIA"
),
ResidenceCode = case_when(
ContinentalRegion == "Africa" ~ 903,
ContinentalRegion == "Asia" ~ 935,
ContinentalRegion == "Europe" ~ 908,
ContinentalRegion == "Latin America and the Caribbean" ~ 904,
ContinentalRegion == "Northern America" ~ 905,
ContinentalRegion == "Oceania" ~ 909
),
LabelR = paste0("<b>Continental Region: </b>", Residence, "<br>",
"<b>Continent: </b>", Continent,"<br>",
"<b>M49 Code: </b>", ContinentCode,"<br>"),
CLRCode = c("#339900", "#FF9900", "#003366", "#FF0000", "#663399", "#FFCC00")) %>%
select(1, 5:6, 2:3, 7:8, 4)
RDATA.CRs <- GeoRDATA.CRs %>% st_drop_geometry()
GeoRDATA.CSRs <- GeoDATA %>%
group_by(SIRegion) %>%
summarize(
SIRegionCode = Mode(SIRegionCode),
Continent = Mode(Continent),
ContinentCode = Mode(ContinentCode),
geometry = st_combine(geometry)) %>%
mutate(
ResidenceCode = case_when(
SIRegionCode == 53 ~ 927,
SIRegionCode == 29 ~ 915,
SIRegionCode == 13 ~ 916,
SIRegionCode == 143 ~ 5500,
SIRegionCode == 14 ~ 910,
SIRegionCode == 30 ~ 906,
SIRegionCode == 151 ~ 923,
SIRegionCode == 54 ~ 928,
SIRegionCode == 57 ~ 954,
SIRegionCode == 17 ~ 911,
SIRegionCode == 15 ~ 912,
SIRegionCode == 21 ~ 905,
SIRegionCode == 154 ~ 924,
SIRegionCode == 61 ~ 957,
SIRegionCode == 5 ~ 931,
SIRegionCode == 35 ~ 920,
SIRegionCode == 18 ~ 913,
SIRegionCode == 34 ~ 5501,
SIRegionCode == 39 ~ 925,
SIRegionCode == 11 ~ 914,
SIRegionCode == 145 ~ 922,
SIRegionCode == 155 ~ 926
),
LabelR = paste0("<b>Continental Sub-Region: </b>", SIRegion, "<br>",
"<b>M49 Code: </b>", SIRegionCode,"<br>"),
# CLRCode => Latin America
# SR <- c("Caribbean", "Central America", "South America")
# colorRampPalette(c("#FF0000", "#660000"))(length(SR))
CLRCode = c("#FFCC00", "#FF0000", "#B20000", "#FFCC99", "#CCFF33",
"#CC9966", "#33CCFF", "#FFFF00", "#FFFF99", "#99CC33",
"#66CC33", "#663399", "#3399CC", "#FFFFCC", "#660000",
"#996600", "#339900", "#FF9900", "#006699", "#336600",
"#FF6600", "#003366")) %>%
select(1:2, 6, 3:4, 7:8, 5)
RDATA.CSRs <- GeoRDATA.CSRs %>% st_drop_geometry()
GeoRDATA.GRSDG <- GeoDATA %>%
group_by(ClasseSDG) %>%
summarize(geometry = st_combine(geometry)) %>%
mutate(ResidenceCode = case_when(
ClasseSDG == "Australia and New Zealand" ~ 927,
ClasseSDG == "Central and Southern Asia" ~ 921,
ClasseSDG == "Eastern and South-Eastern Asia" ~ 1832,
ClasseSDG == "Europe and Northern America" ~ 1829,
ClasseSDG == "Latin America and the Caribbean" ~ 1830,
ClasseSDG == "Northern Africa and Western Asia" ~ 1833,
ClasseSDG == "Oceania*" ~ 1835,
ClasseSDG == "Sub-Saharan Africa" ~ 947,
),
LabelR = paste0("<b>Geographic Region: </b>", ClasseSDG),
CLRCode = c("#FF0000", "#FF6600", "#339900", "#003366",
"#33CCFF", "#FF9900", "#CC0000", "#FF3399")) %>%
select(1, 3:5, 2)
RDATA.GRSDG <- GeoRDATA.GRSDG %>% st_drop_geometry()
GeoRDATA.IncomeLevel <- GeoDATA %>%
group_by(IncomeLevel) %>%
summarize(geometry = st_combine(geometry)) %>%
mutate(
Residence = case_when(
IncomeLevel == "H" ~ "High-Income Countries",
IncomeLevel == "L" ~ "Low-Income Countries",
IncomeLevel == "LM" ~ "Lower-Middle-Income Countries",
IncomeLevel == "UM" ~ "Upper-Middle-Income Countries",
is.na(IncomeLevel) ~ "Not Available"
),
ResidenceCode = case_when(
IncomeLevel == "H" ~ 1503,
IncomeLevel == "L" ~ 1500,
IncomeLevel == "LM" ~ 1501,
IncomeLevel == "UM" ~ 1502,
IncomeLevel == NA ~ NA
),
LabelR = paste0("<b>Income Level (2020): </b>", Residence),
CLRCode = c("#339900", "#663399", "#CC99FF", "#99FF66", "#999999")) %>%
select(3:6, 2) %>%
arrange(desc(ResidenceCode))
RDATA.IncomeLevel <- GeoRDATA.IncomeLevel %>% st_drop_geometry()
CRs <- unique(GeoDATA$ContinentalRegion)
CCodes <- data.frame(Residence = character(), CLRCode = character())
for (CR in CRs) {
if (CR == "Africa") {
C0 <- "#CCFF33"
C1 <- "#336600"
}
else if (CR == "Asia") {
C0 <- c("#FF9900", "#FFCC99")
C1 <- c("#FF6600", "#996600")
L0 <- c(1, 26)
L1 <- c(25, 51)
}
else if (CR == "Europe") {
C0 <- "#33CCFF"
C1 <- "#003366"
}
else if (CR == "Latin America and the Caribbean") {
C0 <- "#FF0000"
C1 <- "#660000"
}
else if (CR == "Northern America") {
C0 <- "#CC99FF"
C1 <- "#663399"
}
else if (CR == "Oceania") {
C0 <- "#FFCC00"
C1 <- "#FFFFCC"
}
Countries <- GeoDATA$Residence[which(GeoDATA$ContinentalRegion == CR & !is.na(GeoDATA$ResidenceCode))[order(GeoDATA$Residence[which(GeoDATA$ContinentalRegion == CR & !is.na(GeoDATA$ResidenceCode))])]]
if (CR == "Asia") {
ColorCodes <- unlist(sapply(1:2,
function(i) colorRampPalette(c(C0[i], C1[i]))(length(Countries[L0[i]:L1[i]]))))
} else {
ColorCodes <- colorRampPalette(c(C0, C1))(length(Countries))
}
CCodes <- rbind(CCodes, data.frame(Residence = Countries, CLRCode = ColorCodes))
}
GeoDATA <- merge(GeoDATA, CCodes, by = "Residence", all.x = TRUE)
GeoDATA$CLRCode <- ifelse(is.na(GeoDATA$CLRCode), "#666666", GeoDATA$CLRCode)
DATA <- GeoDATA %>% st_drop_geometry()5.3 WRITE
# write.xlsx(IMS.DATA.SHARE, "SERVER+/VERSION.1.1/Données/IMS.xlsx")
#
# st_write(GeoDATA, "SERVER+/VERSION.1.1/Données/Geo/Countries/RIMS_WAB_UNSD_WorldBank.geojson",
# driver = "GeoJSON")
# write.xlsx(DATA, "SERVER+/VERSION.1.1/Données/Geo/Countries/DATA.xlsx")
#
# st_write(GeoRDATA.CRs, "SERVER+/VERSION.1.1/Données/Geo/Regions/CRs/GeoRDATA.CRs.geojson",
# driver = "GeoJSON")
# write.xlsx(RDATA.CRs, "SERVER+/VERSION.1.1/Données/Geo/Regions/CRs/RDATA.CRs.xlsx")
#
# st_write(GeoRDATA.CSRs, "SERVER+/VERSION.1.1/Données/Geo/Regions/CSRs/GeoRDATA.CSRs.geojson",
# driver = "GeoJSON")
# write.xlsx(RDATA.CSRs, "SERVER+/VERSION.1.1/Données/Geo/Regions/CSRs/RDATA.CSRs.xlsx")
#
# st_write(GeoRDATA.GRSDG, "SERVER+/VERSION.1.1/Données/Geo/Regions/GRSDG/GeoRDATA.GRSDG.geojson",
# driver = "GeoJSON")
# write.xlsx(RDATA.GRSDG, "SERVER+/VERSION.1.1/Données/Geo/Regions/GRSDG/RDATA.GRSDG.xlsx")
#
# st_write(GeoRDATA.IncomeLevel,
# "SERVER+/VERSION.1.1/Données/Geo/Regions/IncomeLevel/GeoRDATA.IncomeLevel.geojson",
# driver = "GeoJSON")
# write.xlsx(RDATA.IncomeLevel,
# "SERVER+/VERSION.1.1/Données/Geo/Regions/IncomeLevel/RDATA.IncomeLevel.xlsx")